Re: [DNG] Trouble getting copied system to boot
Hendrik Boom schrieb: > On Sun, Dec 17, 2017 at 02:03:52PM +0100, > > It would be something like > > mkdir /ascii/dev > mount --bind /dev /ascii/dev > chroot /ascii > update-initramfs -u > update-grub Yes, that's it, besides that the /dev/ directory should definitely already exist inside your copied (target) root partition. So just bind-mount the active host's (populated) /dev/ there. If you have the (i think it's called) "os-prober" package installed on both systems, you can run update-grub also from outside the chroot jail. Libre grüße, Florian ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble getting copied system to boot
Le 18/12/2017 à 02:29, Hendrik Boom a écrit : On Sun, Dec 17, 2017 at 11:36:19PM +0100, Didier Kryn wrote: Le 17/12/2017 à 19:13, Hendrik Boom a écrit : On Sun, Dec 17, 2017 at 02:03:52PM +0100, Florian Zieboll wrote: On Sat, 16 Dec 2017 21:42:05 -0500 Hendrik Boomwrote: How do I update the initrd? To update the /initrd/initramfs/ of a not-booting system, I chroot into it, (mount the /boot partition) and "bind mount" /dev. Then $ update-initramfs -u and $ update-grub First time I've done something like that in chroot. It would be something like mkdir /ascii/dev mount --bind /dev /ascii/dev chroot /ascii update-initramfs -u update-grub ? /ascii/dev should exist already. You just need to mount it. It doesn't. I suggest also the following: mount -t proc none /ascii/proc mount -t sysfs none /ascii/sys This raises the question: "how did you clone your OS?" Because, if you copied it while /dev, /sys, /proc, and /run where mounted, then you have stored a lot of things that shouldn't be there. Better check and empty those directories on /ascii. I copied them with rsync, using the option that copies one file system only; i.e. does not cross file system boundaries. That's why there is no /ascii/dev -- the original /dev was a mounted file system created by I presume udev. But you need the mountpoints for /dev, /proc, /sys, that is empty directories. Look at what you have in / and check you have all the same directories in /ascii (not /ascii/ascii, of course). The OS is spread in /bin, /boot, /lib, /lib32, /lib64, /etc, /opt, /root, /sbin, /srv, /usr, /var. Some can be empty (/root, /opt, /srv); apt-get dist-upgrade will modify the contents of all these directories. They should, of course not be shared between your two OSes. You can use rsync as you did, for all the directories other than proc, sys, run and dev, after all those which are mountpoints have been mounted. Then mount /ascii/{proc, sys, dev} as said above, and then chroot. And, before booting your new OS, check carefully the fstab. Which is why I thought I had to create the mount point in ascii. Now I suspect I'd also have to create /sys /proc and /run as additional mount points? Or should I start over, booting refracta instead of my new system and using refracta to copy everything when everything isn't the working system? Cloning a running OS is non-trivial. It is easier to clone it when it is not running, for example, plugging the disk in another machine running its own OS, or using a third OS on the same machine, as I understand you propose.ALso, beware that the two OSes should not be mixed. At the maximum, they can share /home, and, even for /home it is dangerous: the applications on the newer OS will update users' config filesthe first time they are used (eg firefox or thunderbird) and make them unusable in the older OS. For what concerns grub, as far as I understand, you should not need to chroot to run update-grub. update-grub is supposed to detect all OSes present on the machine and create a boot menu. Better run it on your main OS. Didier ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble getting copied system to boot
On Sun, Dec 17, 2017 at 07:40:51PM -0600, goli...@dyne.org wrote: > On 2017-12-17 19:29, Hendrik Boom wrote: > > > >Or should I start over, booting refracta instead of my new system and > >using refracta to copy everything when everything isn't the working > >system? > > > > IIUC refractasnapshot only duplicates the running system No No No! I'd use rsync to copy my usual system when it's not running! -- hendrik > > golinux > ___ > Dng mailing list > Dng@lists.dyne.org > https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble getting copied system to boot
On 2017-12-17 19:29, Hendrik Boom wrote: Or should I start over, booting refracta instead of my new system and using refracta to copy everything when everything isn't the working system? IIUC refractasnapshot only duplicates the running system golinux ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble getting copied system to boot
On Sun, Dec 17, 2017 at 11:36:19PM +0100, Didier Kryn wrote: > Le 17/12/2017 à 19:13, Hendrik Boom a écrit : > >On Sun, Dec 17, 2017 at 02:03:52PM +0100, Florian Zieboll wrote: > >>On Sat, 16 Dec 2017 21:42:05 -0500 > >>Hendrik Boomwrote: > >> > >>>How do I update the initrd? > >>To update the /initrd/initramfs/ of a not-booting system, I chroot into > >>it, (mount the /boot partition) and "bind mount" /dev. Then > >> > >>$ update-initramfs -u > >> > >>and > >> > >>$ update-grub > >First time I've done something like that in chroot. > >It would be something like > > > >mkdir /ascii/dev > >mount --bind /dev /ascii/dev > >chroot /ascii > >update-initramfs -u > >update-grub > > > >? > > > > /ascii/dev should exist already. You just need to mount it. It doesn't. > I suggest also the following: > mount -t proc none /ascii/proc > mount -t sysfs none /ascii/sys > > This raises the question: "how did you clone your OS?" Because, if you > copied it while /dev, /sys, /proc, and /run where mounted, then you have > stored a lot of things that shouldn't be there. Better check and empty those > directories on /ascii. I copied them with rsync, using the option that copies one file system only; i.e. does not cross file system boundaries. That's why there is no /ascii/dev -- the original /dev was a mounted file system created by I presume udev. Which is why I thought I had to create the mount point in ascii. Now I suspect I'd also have to create /sys /proc and /run as additional mount points? Or should I start over, booting refracta instead of my new system and using refracta to copy everything when everything isn't the working system? -- hendrik kk ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble getting copied system to boot
Le 17/12/2017 à 19:13, Hendrik Boom a écrit : On Sun, Dec 17, 2017 at 02:03:52PM +0100, Florian Zieboll wrote: On Sat, 16 Dec 2017 21:42:05 -0500 Hendrik Boomwrote: How do I update the initrd? To update the /initrd/initramfs/ of a not-booting system, I chroot into it, (mount the /boot partition) and "bind mount" /dev. Then $ update-initramfs -u and $ update-grub First time I've done something like that in chroot. It would be something like mkdir /ascii/dev mount --bind /dev /ascii/dev chroot /ascii update-initramfs -u update-grub ? /ascii/dev should exist already. You just need to mount it. I suggest also the following: mount -t proc none /ascii/proc mount -t sysfs none /ascii/sys This raises the question: "how did you clone your OS?" Because, if you copied it while /dev, /sys, /proc, and /run where mounted, then you have stored a lot of things that shouldn't be there. Better check and empty those directories on /ascii. Didier ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble getting copied system to boot
On Sun, Dec 17, 2017 at 02:03:52PM +0100, Florian Zieboll wrote: > On Sat, 16 Dec 2017 21:42:05 -0500 > Hendrik Boomwrote: > > > How do I update the initrd? > > To update the /initrd/initramfs/ of a not-booting system, I chroot into > it, (mount the /boot partition) and "bind mount" /dev. Then > > $ update-initramfs -u > > and > > $ update-grub First time I've done something like that in chroot. It would be something like mkdir /ascii/dev mount --bind /dev /ascii/dev chroot /ascii update-initramfs -u update-grub ? -- hendrik > > Libre Grüße, > > Florian > > > ___ > Dng mailing list > Dng@lists.dyne.org > https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble getting copied system to boot
On Sun, Dec 17, 2017 at 12:37:03PM +0100, Didier Kryn wrote: > Le 16/12/2017 à 23:55, Hendrik Boom a écrit : > >I m copying my entire Devuan system to new partitions on the same computer. > >but I can't get the copy system to boot. > > > >I plan to upgrade the copy to ascii, keeping the old system in dual-boot > >scenario just in case I'm not expecting problems, but I've had them on > >Debian upgrades years ago, and they sometimes arise from causes that have > >nothing to do with the quality of the old and new systems. > > > >The messge I get is: > > > >vmlinuz-3.16.0-4-686-pae not found > >you need to load the kernel first > > > >But vmlinuz-3.16.0-4-686-pae *is* the kernel. I don't know how to load the > >ernel before it can find the kernel. > > > >*** > > > >Here's how I got to this point. > > > >/boot and /ascii/boot (I'll use the names these partitions have on the old > >system) are partitions of their own on a DOS-style partitioned disk, as is > >swap. > > > >/dev/sda6 on /boot type ext4 (rw,relatime,data=ordered) > >/dev/sda7 on /ascii/boot type ext4 (rw,relatime,data=ordered > > > >oot@notlookedfor:/home/hendrik# swapon > >NAME TYPE SIZE USED PRIO > >/dev/sda5 partition 4.7G 0B -1 > >root@notlookedfor:/home/hendrik# > > > >The other partitions are within LVM. > > > >/dev/mapper/jessie-devuan--home on /home type ext4 (rw,relatime,data=ordered) > >/dev/mapper/jessie-tmp on /tmp type ext4 (rw,relatime,data=ordered) > >/dev/sda6 on /boot type ext4 (rw,relatime,data=ordered) > >/dev/mapper/jessie-devuan--usr on /usr type ext4 (rw,relatime,data=ordered) > >/dev/mapper/jessie-devuan--var on /var type ext4 (rw,relatime,data=ordered) > >/dev/sda7 on /ascii/boot type ext4 (rw,relatime,data=ordered) > >/dev/mapper/jessie-ascii--root on /ascii/root type ext4 > >(rw,relatime,errors=remount-ro,data=ordered) > >/dev/mapper/jessie-ascii--usr on /ascii/usr type ext4 > >(rw,relatime,data=ordered) > >/dev/mapper/jessie-ascii--var on /ascii/var type ext4 > >(rw,relatime,data=ordered) > > > >The partitions with devuan in their mapper names belong to the jessie > >system; > >those with ascii in theor names are intended for the future ascii system, > >and currently are a copy of the ones for jessie, exceot of course that > >/jessie/root/etc/fstab (ascii's fstab) hace had the partition names swapped > >around, so that for example /boot there is ascii's /boot instead of > >jessie's, and the old systems partitions are known as /jessie/boot, etc. > > > >Just before today's tests, I used rsync to synchronise the two systems, and > >then copied the modufied fstab to the ascii system, just to make sure it was > >the new one instead of the rsync'd one. > > > >Then reboot. It offered me a choice of systems to boot; I chose the new > >one, and got the error message I presented above. > > > >vmlinuz-3.16.0-4-686-pae not found > >you need to load the kernel first > > > >So ... where do I start to look for the problem. How far along in the boot > >process is the system component that produces the above message? Are there > >any usable diagnostic tools at this point in the boot process? > > > >Yes, that file *is* present in /ascii/boot. > > > >root@notlookedfor:/home/hendrik# ls /ascii/boot > >config-3.16.0-4-686-pae lost+found > >grub System.map-3.16.0-4-686-pae > >initrd.img-3.16.0-4-686-pae vmlinuz-3.16.0-4-686-pae > >root@notlookedfor:/home/hendrik# > > I guess your two root partitions must have different uuids and Grub must > know it to choose. > > Thats the problem of specifying boot partition by uuid. I like using > uuids in fstab to keep each OS consistent, but for the root partition passed > to Grub, I personnaly think device name is just more convenient. I guess > this consideration doesn't solve your problem though :-( > > Check also the uuids in your new fstab is fine and won't mix with the > other OS. Both OS's mount the same partitions, with the same UUIDs. Only the mount points are different. I makes it easier to compare them. -- hendrik ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble getting copied system to boot
On Sat, 16 Dec 2017 21:42:05 -0500 Hendrik Boomwrote: > How do I update the initrd? To update the /initrd/initramfs/ of a not-booting system, I chroot into it, (mount the /boot partition) and "bind mount" /dev. Then $ update-initramfs -u and $ update-grub Libre Grüße, Florian pgpQT_qrZd2Cz.pgp Description: OpenPGP digital signature ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble getting copied system to boot
Le 16/12/2017 à 23:55, Hendrik Boom a écrit : I m copying my entire Devuan system to new partitions on the same computer. but I can't get the copy system to boot. I plan to upgrade the copy to ascii, keeping the old system in dual-boot scenario just in case I'm not expecting problems, but I've had them on Debian upgrades years ago, and they sometimes arise from causes that have nothing to do with the quality of the old and new systems. The messge I get is: vmlinuz-3.16.0-4-686-pae not found you need to load the kernel first But vmlinuz-3.16.0-4-686-pae *is* the kernel. I don't know how to load the ernel before it can find the kernel. *** Here's how I got to this point. /boot and /ascii/boot (I'll use the names these partitions have on the old system) are partitions of their own on a DOS-style partitioned disk, as is swap. /dev/sda6 on /boot type ext4 (rw,relatime,data=ordered) /dev/sda7 on /ascii/boot type ext4 (rw,relatime,data=ordered oot@notlookedfor:/home/hendrik# swapon NAME TYPE SIZE USED PRIO /dev/sda5 partition 4.7G 0B -1 root@notlookedfor:/home/hendrik# The other partitions are within LVM. /dev/mapper/jessie-devuan--home on /home type ext4 (rw,relatime,data=ordered) /dev/mapper/jessie-tmp on /tmp type ext4 (rw,relatime,data=ordered) /dev/sda6 on /boot type ext4 (rw,relatime,data=ordered) /dev/mapper/jessie-devuan--usr on /usr type ext4 (rw,relatime,data=ordered) /dev/mapper/jessie-devuan--var on /var type ext4 (rw,relatime,data=ordered) /dev/sda7 on /ascii/boot type ext4 (rw,relatime,data=ordered) /dev/mapper/jessie-ascii--root on /ascii/root type ext4 (rw,relatime,errors=remount-ro,data=ordered) /dev/mapper/jessie-ascii--usr on /ascii/usr type ext4 (rw,relatime,data=ordered) /dev/mapper/jessie-ascii--var on /ascii/var type ext4 (rw,relatime,data=ordered) The partitions with devuan in their mapper names belong to the jessie system; those with ascii in theor names are intended for the future ascii system, and currently are a copy of the ones for jessie, exceot of course that /jessie/root/etc/fstab (ascii's fstab) hace had the partition names swapped around, so that for example /boot there is ascii's /boot instead of jessie's, and the old systems partitions are known as /jessie/boot, etc. Just before today's tests, I used rsync to synchronise the two systems, and then copied the modufied fstab to the ascii system, just to make sure it was the new one instead of the rsync'd one. Then reboot. It offered me a choice of systems to boot; I chose the new one, and got the error message I presented above. vmlinuz-3.16.0-4-686-pae not found you need to load the kernel first So ... where do I start to look for the problem. How far along in the boot process is the system component that produces the above message? Are there any usable diagnostic tools at this point in the boot process? Yes, that file *is* present in /ascii/boot. root@notlookedfor:/home/hendrik# ls /ascii/boot config-3.16.0-4-686-pae lost+found grub System.map-3.16.0-4-686-pae initrd.img-3.16.0-4-686-pae vmlinuz-3.16.0-4-686-pae root@notlookedfor:/home/hendrik# I guess your two root partitions must have different uuids and Grub must know it to choose. Thats the problem of specifying boot partition by uuid. I like using uuids in fstab to keep each OS consistent, but for the root partition passed to Grub, I personnaly think device name is just more convenient. I guess this consideration doesn't solve your problem though :-( Check also the uuids in your new fstab is fine and won't mix with the other OS. Didier Didier ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble getting copied system to boot
Am 2017-12-17 03:43, schrieb Hendrik Boom: Yes. Hex digit by hex digit. Of course I might have done it wrong. I will check again. Tomorrow. When I'm awake again. The easiest way solving boot problems is most often the boot repair disc. https://sourceforge.net/p/boot-repair/home/Home/ Jochen ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble getting copied system to boot
On Sun, Dec 17, 2017 at 12:09:46AM +0100, Dr. Nikolaus Klepp wrote: > Am Samstag, 16. Dezember 2017 schrieb Hendrik Boom: > > I m copying my entire Devuan system to new partitions on the same > > computer. but I can't get the copy system to boot. > > > > I plan to upgrade the copy to ascii, keeping the old system in dual-boot > > scenario just in case I'm not expecting problems, but I've had them on > > Debian upgrades years ago, and they sometimes arise from causes that have > > nothing to do with the quality of the old and new systems. > > > > The messge I get is: > > > > vmlinuz-3.16.0-4-686-pae not found > > you need to load the kernel first > > > > But vmlinuz-3.16.0-4-686-pae *is* the kernel. I don't know how to load the > > ernel before it can find the kernel. > > > > *** > > > > Here's how I got to this point. > > > > /boot and /ascii/boot (I'll use the names these partitions have on the old > > system) are partitions of their own on a DOS-style partitioned disk, as is > > swap. > > > > /dev/sda6 on /boot type ext4 (rw,relatime,data=ordered) > > /dev/sda7 on /ascii/boot type ext4 (rw,relatime,data=ordered > > > > oot@notlookedfor:/home/hendrik# swapon > > NAME TYPE SIZE USED PRIO > > /dev/sda5 partition 4.7G 0B -1 > > root@notlookedfor:/home/hendrik# > > > > The other partitions are within LVM. > > > > /dev/mapper/jessie-devuan--home on /home type ext4 > > (rw,relatime,data=ordered) > > /dev/mapper/jessie-tmp on /tmp type ext4 (rw,relatime,data=ordered) > > /dev/sda6 on /boot type ext4 (rw,relatime,data=ordered) > > /dev/mapper/jessie-devuan--usr on /usr type ext4 (rw,relatime,data=ordered) > > /dev/mapper/jessie-devuan--var on /var type ext4 (rw,relatime,data=ordered) > > /dev/sda7 on /ascii/boot type ext4 (rw,relatime,data=ordered) > > /dev/mapper/jessie-ascii--root on /ascii/root type ext4 > > (rw,relatime,errors=remount-ro,data=ordered) > > /dev/mapper/jessie-ascii--usr on /ascii/usr type ext4 > > (rw,relatime,data=ordered) > > /dev/mapper/jessie-ascii--var on /ascii/var type ext4 > > (rw,relatime,data=ordered) > > > > The partitions with devuan in their mapper names belong to the jessie > > system; > > those with ascii in theor names are intended for the future ascii system, > > and currently are a copy of the ones for jessie, exceot of course that > > /jessie/root/etc/fstab (ascii's fstab) hace had the partition names swapped > > around, so that for example /boot there is ascii's /boot instead of > > jessie's, and the old systems partitions are known as /jessie/boot, etc. > > > > Just before today's tests, I used rsync to synchronise the two systems, and > > then copied the modufied fstab to the ascii system, just to make sure it > > was the new one instead of the rsync'd one. > > > > Then reboot. It offered me a choice of systems to boot; I chose the new > > one, and got the error message I presented above. > > > > vmlinuz-3.16.0-4-686-pae not found > > you need to load the kernel first > > > > So ... where do I start to look for the problem. How far along in the boot > > process is the system component that produces the above message? Are there > > any usable diagnostic tools at this point in the boot process? > > > > Yes, that file *is* present in /ascii/boot. > > > > root@notlookedfor:/home/hendrik# ls /ascii/boot > > config-3.16.0-4-686-pae lost+found > > grub System.map-3.16.0-4-686-pae > > initrd.img-3.16.0-4-686-pae vmlinuz-3.16.0-4-686-pae > > root@notlookedfor:/home/hendrik# > > > > -- hendrik > > you have checked if the UUID of the new root filesystem is the one that GRUB > uses and you have run "update-grub" on the new partition? Yes. Hex digit by hex digit. Of course I might have done it wrong. I will check again. Tomorrow. When I'm awake again. -- hendrik > > Nik > > > > -- > Please do not email me anything that you are not comfortable also sharing > with the NSA, CIA ... > ___ > Dng mailing list > Dng@lists.dyne.org > https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble getting copied system to boot
On Sun, Dec 17, 2017 at 12:10:09AM +0100, Florian Zieboll wrote: > Am 16. Dezember 2017 23:55:46 MEZ schrieb Hendrik Boom >: > > > vmlinuz-3.16.0-4-686-pae not found > > you need to load the kernel first > > hallo hendrik, > > reading over your mail, i couldn't find that you updated the initrd (chroot), > which feels like sound reason for this error. I copied the initrd along with everything else. Does it have dependencies on the specific copies of the file system being used? How do I update the initrd? -- hendrik > > libre grüße, Free greetings! -- hendrik > > florian > > -- > > [message sent mobile] > ___ > Dng mailing list > Dng@lists.dyne.org > https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng